This is worthy of a +1 or moderator vote up.
One of the driving forces for video is the media. They lust after any content that they can get and
rebroadcast.

Their greed could make you a buck, Copyright any home video and communicate it to any responsible agency
with a reminder that it is copyright but never to a news outlet. Should a media outlet use it without permission
go after them all. As a minimum 4x the advertising fee and also salary of the personalities and production staff
involved.

If you are in an aircraft situation... take a phone video -- it might survive.

The damaged data recording boxes are the detail in this that is missing.

Redundant durable data recording boxes in contrast to data collection boxes seem to be the next improvement.
Next, it seems to me that the recovery of these boxes and their data is a weak link in crash analysis.

Modern GPS and multi axis sensors kin to the ones in our phones are inexpensive so the sensor set
can be augmented and improved without touching flight or engine control systems.

Data links to ground recording stations is the next obvious improvement.
Air to air grid style networking can bridge vast reaches over oceans without
connecting to expensive satellite systems. Routine data is modest. Store and forward
strategies can include automated data dump at the ground terminal can minimize live data
links. i.e. fleet data collection aircraft to aircraft for long haul transit.

Cabin alarms can trigger wider air to air transfers and perhaps satellite link dumps.
The big red button could also trigger video transfers and archives but that is simply
another sensor. Video through chewing gum or cabin cleaner goo is a problem.

Summary: multiple data storage boxes with augmented sensors that do not touch flight
control or engine systems are needed. Aircraft to aircraft RF data links can eliminate box recovery urgency.
Note line of sight RF at 25-40,000 feet is a long way. Aircraft can be routed in pairs separated
by 50 miles... even.

This is interesting -- this can allow distribution of
clocks and signals in ways that are possibly immune from
a number of types of signal problems.

It eliminates issues of DC bias and the need for long balanced
transmission line pairs on the PWB. This can allow wider
and thus 2x faster data paths. It eliminates capacitance loading
and associated time delay impact on long transmission lines
perhaps improving signal edges at distance on a PWB.
It may allow signaling between positive and negative logic families
and signaling between different voltage/ power domains.

Time will tell... but just as Swatch was able to eliminate case
packaging issues this could allow the construction of a module
that would make the old IBM ECL based TCM fade into greater
obscurity.

I am one of those gray haired fuddy duddies.
I wonder if I was to hire a wet nurse for myself and if
I was to then go back to school would I become a gazillionaire?

Might be an interesting experiment. ;)

This seems self serving.

Since the set of Mockrosoft products serve as host to the largest
collection of hacked robot farms out there this is interesting. Hardening
their server and other products seems too hard for MS. It seems to me that this may
prove to be the single most cost effective strategy there is to reduce
the size of distributed attack farms. That alone would make their server
products measurably better to customers. It would allow sites
to maintain desirable uptime and availability numbers.

It also reduces the impact on software engineering in Redmond because
this makes is easier to slowly walk away from previous Windowz versions.

Without knowing the truth, I would assert their cash flow is not dominated by selling updates, it is
Office and new hardware tax.

It may also enable improved markets for new Office products for Asian languages.
Back porting and compatibility in Office 2xxx-new is baggage that might
be left behind.

It is a big bet that Win-10 will run well enough on the older hardware
and a big bet on the quality of the release.

It could pay off...
It could just make Linux+GNU a better choice.

This is important....
I should add that plate readers allow this revenue model to be optimized.
Isolation of citation income value data is clearly needed.

Those that see this as a good thing need to be monitored with care.
The only good news is the RICH do have long arms and big legal sticks.
They also finance individual election programs and abusers might find themselves
paying fines for walking 6 mph in a 3 mph walking zone inside a fenced
exercise yard.

Some places do consider safety a goal.

If safety is a goal you make the fine large enough to cause the offender some
financial pain.

But wait if you earn millions and the next guy makes hundreds this very much
levels the field of pain.

One troubling problem with traffic fines is the sum and the finances of most
receiving the fine are upside down and the disenfranchised have insufficient
resources to fight the systematic problems with many traffic laws and their
enforcement.

A fine this large justifies legal attention and a side effect is that the judge
will see improved defense and review of the law.

Yes, I think it is outrageous but then I do not earn millions.

My gut instinct is that this is the kernel of some improvement
to the gross abuses that some traffic enforcement programs finance
themselves with.

The next obvious to me abusive process is the photo citations and escalating
fail to appear bench warrant processes farmed out to civilian contracts.
The citation is often waved for want of evidence but the fail to appear
is self-proving and can be isolated from the initial systematic fraud.

How much of this feedbag is from apple.com or another
competitor?

My CP/M system is still giving me fine service air-gapped
from the universe.

More importantly how is this pile broken down.
Some hate any change... bucket A.
Some find broken stuff... bucket B (B as in badly broken bozo)
Some want their personal change ... bucket C.
Some found dumb stuff ... bucket D.

The report appears innocuous but is also justification for more air power.

Hidden from us is the effect of flattening hot crime spots and dispersing crime more evenly across the area. Short term reduction of crime in hot spots seems very real but would identify the hot spot and move crime to cooler spots.

It does little to solve the social and economic wreckage in many neighborhoods that makes crime the most profitable activity.

With deep database background searches no past criminal can get an "interesting" or well paying job. With 20-40% of the mail population in some areas there are rare honest jobs.

The multi million budget for one helicopter would better be spent on solving social problems. This is harder to do than I like but it needs to be done.

We are making some improvements with the decriminalization of marijuana but have failed to discuss a need to expunge non violent non repeat crimes from public employment screening. Simply financing tattoo removal would help some individuals.

http://en.wikipedia.org/wiki/L...

Nothing against the unit but the budget area and population make me wonder if more is justified.

So when someone dies how does that person make an entry in the SSN database?

Auto deposit matched with an auto withdrawal by Medicare.
Life insurance automatic payment is the insurance company going to care.

Election boards.. love to have a fat census. They love to have a fat get out the vote budget.
Clearly they need more money because 99.8% of all the individuals over 100 fail to
show up at the polls.

Repetitive stress will be the problem in many....
The deletion window has opened and the bits are flying
away.

Now if only Hillary had her server in Dutch jurisdiction....

The Gentoo crowd had a hoop or two to jump through to get from one version of gcc to another way back when.
Compilers and ABI designs are important -- the fuzzy rules for ARM ABIs is holding ARM back for some.

Linus may be correct from where he sits. A lot of where he sits is atop a massive
pile of C and history written in C going back to Minix and other versions of Unix -- all of which
were built with and on C.

Some of the microkernel designs could have a leg up and the close to hardware
bits could be isolated from upper layers that could be crafted in another language.

So if you want to start over and build from the ground up... who knows.
But today "C" is the anchor for the pile of stuff that Linus sits on.

Hmmm this sandbox strategy is used by Firefox and many more tools.
As more and more tools move to threads this ability
to sync them will gain traction.

My guess is there is a window of risk that needs to be closed before it surfaces
as a bug or exploit. All in all this sandbox stuff is new but interesting as heck.
There are stronger models but this is an improvement especially when RAM is
limited -- (tablets and phones).